Rethinking ethnicity, 2nd Edition by Richard Jenkins

72 Rethinking Ethnicity

The first is that indispensable adjunct to the planning processes of modern government, the census of population (Cohn 1988: 224–54; Kertzer and Arel 2002; Sundar 2000). In order to gather meaningful data, population categories must first be defined. Among these are ethnic (or ‘racial’) categories (Lee 1993). Thus they become established in official discourses, discourses which are powerfully constitutive of social reality through public rhetoric, the formulation of policy, the targeting of resources and social control measures. The census categories may be used to collect and construct other statistics – on crime, unemployment, health – which also feed into policy formulation and implementation. These official categorizations are not necessarily directly reappropriated as selfor group identifications. History, context, and, not least, the content and the consequences of the categorization, all matter. For example, the diverse membership of the ‘Asian-American’ category, a product of the US Census and the classificatory needs of affirmative action, have yet to be welded into anything other than a contingent, instrumental, categorical collectivity. A ‘pan-ethnic’ group identification remains elusive (LeEspiritu 1992). In Britain, the category ‘black’, favoured by some ethnic minority activists as a response to a common experience of racism, is in tension with a range of cultural and class differentiations which continue to render it a highly contested identification (Modood 1992). However, because these ethnic categories – or, indeed, the collection of the information at all – may, for a variety of reasons, be contested by the sections of the population to whom they are applied (see also Blum 2002), and the state may use them to justify consequential policy in areas such as immigration control, internal policing and collective resource allocation (Anderson and Fienberg 2000; Bhat et al. 1988), even census categories cannot be said to be completely irrelevant to group identifications. Their contestation and consequences may contribute, directly and indirectly, to selfand group identification.

The second point is that census and other statistics are collective and reasonably anonymous; they do not directly categorize named individuals (although they may feed into administrative allocation). Indeed, not many states categorize their individual citizens in this fashion. Minor examples of such an approach include the compulsory categorization and recording, for purposes of affirmative action monitoring, of the ethnic identity of employees in many US Federal, state and privatesector organizations, or the identification of ‘nationality’ – denoting Republic of origin – on the pre-1989 USSR internal passport. Other states, however, most notably Nazi Germany and the Republic of South Africa, having defined citizenship in terms of ‘race’ – a crucial escalatory move in ethnic categorization – have gone on, necessarily, to develop far-reaching systems of ‘racial’ categorization, governing every aspect of life (and death). The role of ‘expert’ knowledge in the construction of those systems and their bureaucratized administration is a grim reminder of the repressive possibilities inherent in administrative allocation (Bauman 1989; Burleigh and Wippermann 1991; Müller-Hill 1988).

This lamentable history brings me, finally, to science. A telling presence in many of the contexts of social categorization discussed so far, it differs from healing and medicine, for example, in at least two respects that are relevant here. It is by definition (or at least in terms of its own narratives) a wholly formalized – and rational or disinterested – ensemble of discourses and practices, and it is in principle concerned more with general categories than individuals.

Categorization and Power 73

In fact, science as applied to humans is one of the most pervasive and significant contexts of collective social categorization, at one historical period or another generating and elaborating the following powerful collective identifications: ‘race’; normality, subnormality and abnormality; disability; social class; deviance; a range of mental illnesses and syndromes; a range of physical illnesses and syndromes; sexual orientation; and so on (Hacking 1990; Jenkins 1998a, 1998b). All of these become consequential – and typically far from disinterested – when turned into testable criteria of individual categorization and/or intervention (de Swaan 1990; Hanson 1993). Ethnic and ‘racial’ categorizations continue to figure in scientific practice, and controversies, in fields such as epidemiology (e.g. Kaplan and Bennett 2003; Kaufman and Cooper 2001).

There are many other contexts of ethnic – and indeed other – categorization to which research attention should be directed: friendship relations, education, and religion are among the possibilities. Some of the contexts I have discussed are themselves so broad – administrative allocation, for example – that they could usefully be disaggregated for the purposes of empirical research. A number of points should be remembered in any consideration of the research issues raised by this section:

•First, although my own research background may have encouraged me to draw upon illustrative material from complex, urbanized societies (and disproportionately from Europe and North America), this schematization of ‘contexts of ethnic categorization’ is intended to have the widest possible application. Even so, it is true that practices such as administrative allocation, social policy, official classification and science are deeply implicated in modernity. However, since modernity is now at least the medium-range context of all human life, no matter how isolated or apparently ‘traditional’, and states and their institutions of administration and classification are historically very long established in many parts of the world, this may not place too many constraints on the general applicability of the scheme.

•Second, categorization is unavoidable in our efforts to know and understand the human world, and in all social identification. Thus, although it may be defensible to assume, as I have done here, that ethnic categorization is disproportionately likely to be negative in its consequences (and possibly intent), there is no necessary equation between stigmatization or oppression and categorization in general. Categorization can be neutral, and it can be positive and valorizing (Jenkins 2008: Chapter 8).

•Third, the different social contexts of categorization should not be understood as separate or hierarchical ‘levels’. They overlap systematically in complex and interesting ways: routine public interaction, for example, takes place within a number of the other social contexts mentioned. A similar point could be made about informal social groups, which exist within (and contribute towards the organization of) many different contexts. And so on. One important research task is to problematize and examine this mutual reinforcement.

•Fourth, the mutual reinforcement of categorization in overlapping social contexts – what once might have been called over-determination – contributes to the consequential power and resilience of particular categorizations. Categorizations such as ethnicity or ‘race’, which are often significant in many

74 Rethinking Ethnicity

social settings, may thus harden up into identifications that are robust and resistant to change (cf. Ruane and Todd 2004). This is of great importance in attempting to understand the obstinacy of racism, for example, and, as discussed in Chapter 4, the fierce persistence of ethnic attachments in some settings, even when they might not appear, to the outside eye, to be in the ‘best’ economic or political interests of those who identify with them.

•Finally, the dialectical implication in each other of internal and external definition, group identification and categorization, should not be forgotten. Although the discussion in this section has largely been about categorization, the social contexts identified are also significant in processes of internal definition. Given the interpenetration of the two, how could it be otherwise?

Categorization and power

Categorization contributes to group identity in various ways. There is, for example, something which might be referred to as ‘internalization’: the categorized group is exposed to the terms in which another group defines it and assimilates that categorization, in whole or in part, into its own identity. Put baldly, however, the suggestion seems to beg more questions than it answers: why should the external definition be internalized, for example, and how does it happen? At least five possible scenarios suggest themselves.

In the first, the external categorization might be more or less the same as an aspect of existing group identity, in which case they will simply reinforce each other. It seems altogether plausible, in fact, that some degree of external reinforcement or validation is crucial to the successful maintenance of internal (group) definitions. Similarly, categorization may be less likely to ‘stick’ where it is markedly at odds with existing boundaries and identifications.

Second, there is the incremental cultural change that is likely to be a product of any long-standing but relatively harmonious inter-ethnic contact. The ethnic boundary is osmotic, and not just in terms of personnel: languages and cultures may interact too, and in the process identities are likely to be affected. We, for example, may gradually and imperceptibly come to define ourselves somewhat differently in the light of how they appear to define us, and how they treat us (and, equally likely, vice versa).

In the third place, the external category is produced by Others who possess, in the eyes of the original group, the legitimate authority to categorize them, because of their superior ritual status, knowledge, or whatever. Such a situation implies greater social than cultural differentiation – if a distinction posed in these terms is admissible (Jenkins 2002a: 39–62) – inasmuch as legitimate authority necessarily requires at least a minimal degree of shared participation in values or cosmology.

Fourth is a simpler case, or at least a cruder one: external categorization is imposed by the use of physical force or its threat, i.e. the exercise of power. The experience of violation may become integral to group identification. The categorized, without the physical capacity to resist the carrying of identity cards, the wearing of armbands, or whatever more subtle devices of identification and

Categorization and Power 75

stigmatization might be deployed, may, in time, come to see themselves in the language and categories of the oppressor. They are certainly likely to behave in an appropriate manner.

Finally, there are the oppressed who resist, who reject imposed boundaries and/or their categorical content(s). However, the very act of defying categorization, of striving for autonomous self-identification, is, of course, an effect of being categorized in the first place. The rejected external definition is internalized but, paradoxically, as the focus of denial.

In these five possibilities, a distinction between power and legitimate authority is apparent. However, the contribution of categorization to group identity depends upon more than ‘internalization’. The capacity of one group of people to effectively define or constitute the conditions of existence experienced by Others is enormously important in the internal–external dialectic of collective identification (something similar, at the individual level, was discussed above, in the context of labelling). Categorization is consequential, and it is in the consequences that it may be most effective. This calls up the distinction I have proposed between the virtual and nominal dimensions of ethnic identification, and emphasizes the virtual.

To revisit an example given earlier, the categorization in particular ways of the Mapuche by the Spanish and white Chileans must be expected to have influenced the behaviour of the Chileans towards the Mapuche: ‘native policy’ as Stuchlik puts it. The Chileans, like the Spanish before them, because of their eventual monopolization of violence in the local context and their consequent control of resources, were – and, indeed, still are – in a position to make their categorization of the Mapuche count disproportionately in the social construction of Mapuche life. And the internal–external dialectic can be seen at work: resistance to the Spanish was, for example, an important factor creating internal group identification in the shape of military alliances, and generating, between the sixteenth and nineteenth centuries, a wider collectivity of ‘the Mapuche’ out of networks of small, mutually related kin groups (Stuchlik 1976: 15). In many senses, therefore, what it means to be a ‘Mapuche’ is in part a consequence of what the Spanish or the Chileans have made it mean.

The effective categorization of a group of people by a more powerful Other is not, therefore, ‘just’ a matter of classification (if, indeed, there is any such thing). It is necessarily an intervention in that group’s world that will, to an extent and in ways that are a function of the specifics of the situation, alter that world and the experience of living in it. Just as the Chileans have the capacity to constitute, in part, the experience of ‘being a Mapuche’, so, for example, employment recruiters in Britain – who are still typically white – contribute to the social constitution of the collective experience of growing up as a member of a black ethnic minority.

To return to Barth – and Weber – here one sees the unintended consequences of action at work. It is partly in the cumulative mutual reinforcement of these unintended consequences that the patterns of history, both in the present and as a framework of constraint and possibility for future generations, are produced. Since ‘culture’ is a matter of everyday life and its exigencies, the power of others to constitute the experience of daily living is a further important contribution of categorization to group identity. It is also powerful support for the view of ethnic identification – as everyday practice and historical process – that I have developed here.

76 Rethinking Ethnicity

One of my starting points is that ethnicity should be conceptualized within a theoretical framework that allows for its integration into the topic of ‘identity’ in general. In its turn, however, identity must be constructed as a proper subject for theorization in a manner that allows the inclusion of individual and collective identities within a unified analytical framework (Jenkins 2008). Even the most private of identities is not imaginable as anything other than the product of a socialized consciousness and a social situation. Even the most collective of identities must, in some sense, exist in the awareness of individual actors.

Viewed in the abstract, identity – whether ethnic or otherwise – can be understood as two interacting but independent entailments: a name (the nominal) and an experience (the virtual). The latter is, in a sense, what the name means; this is primarily a matter of its consequences for those who bear it, and can change while the nominal identity remains the same (and vice versa). The nominal and the virtual unite in the ongoing production and reproduction of identity and its boundaries. The nominal–virtual distinction recognizes that ethnic identities – and, indeed, all identities – are practical accomplishments rather than static forms. They are immanently, although not necessarily, variable. As a social form – in the constitution of a historically enduring ethnic group, for example – ethnicity is always potentially variable, nominally and virtually.

In the practical accomplishment of identity, two mutually interdependent but theoretically distinct processes are at work: internal definition and external definition. These operate differently in the individual, interactional and institutional orders. Analyses of ethnicity – particularly within social anthropology – have, however, emphasized internal definition and group identification at the expense of external definition and social categorization. This is more than just a matter of empirical neglect. It is actually a lopsided understanding. Identification is always the practical product of the interaction of ongoing processes of internal and external definition. One cannot be understood in isolation from the other.

A concern with external definition and categorization demands, finally, that analyses of ethnicity should focus on power as well as on authority, and on the manner in which different modes of domination are implicated in the social construction of ethnic and other identities. This is not an original observation. But if we do not do this, the result is likely to be a model of ethnicity that is as trivial as it is one-sided. Unless we can construct an understanding of ethnicity that addresses all of ethnicity’s facets and manifestations – from the celebratory communality of belonging to the final awful moment of genocide – we will have failed not only ourselves, but also the people about whom we write.

6 Ideologies of Identification

In Chapter 4 I suggested that ethnicity is a ubiquitous mode of social identification, and that homologous phenomena such as racism and nationalism can be understood as historically specific allotropes or versions of the wider principle of ethnic affiliation and classification. This chapter will explore these arguments in greater detail and will critically examine some current conceptualizations of ethnicity, racism and nationalism, and the relationships between them.

‘Race’ is typically differentiated from ethnicity in terms of a contrast between physical and cultural differences (Banton 1988: 1–15). Whether explicit or implicit, and however formulated, this basic model underlies most social science discussions of the topic: ‘racial’ differentiation is something – the contentious issue is what – to do with physical differences between people. However, the ‘physical differences’ with which we are concerned in matters of ‘race’ are only differences that make a difference because they are culturally or socially signified as such (Wade 1993, 2002). There is, therefore, nothing ‘objective’ about ‘race’; hence the need felt by many of us to continue to hedge the word around with scare quotes.

Although it is possible to define these concepts in the above general manner, the relationship between ‘race’ and ethnicity is something about which there is little consensus and a good deal of touchiness. There may not even always be agreement that the distinction is, in itself, valid or important. In previous chapters I referred to Sandra Wallman’s argument (1986: 229) that phenotype or physical appearance – and this is what is generally meant by ‘race’ – is no more than ‘one element in the repertoire of [ethnic] boundary markers’. A perspective similar to Wallman’s in many respects, although arrived at from a very different direction – ‘a multiculturalist context that seeks to redress racial disadvantage by providing for ethnic difference’ – is offered by Floya Anthias, who argues ‘that “race” categories belong to the more encompassing category of ethnic collectivity’ (1992: 421). ‘Race’, she suggests, is simply one of the ways in which ethnic boundaries are constructed.

There is much to be said for arguments that recognize a systematic relationship between ethnicity and ‘race’; indeed, this book must be counted among them. However, the versions offered by Wallman or Anthias are unconvincing, for several reasons which I have discussed elsewhere. First, while ‘ethnic’ social relations are not necessarily hierarchical, exploitative and conflictual, ‘race relations’ would certainly appear to be. Although ethnic boundaries often involve relations of power, and social categorization is inherent to the internal–external dialectic of ethnic identification, hierarchical difference is not definitive of ethnic relations. In other words, it is possible to imagine – and, in fact, to document empirically – ethnic inter-group relations in which the groups involved neither look up to nor down on each other. Second, leaving power to one side and simply focusing on process, ‘race’, unlike ethnicity, seems to be much more a matter of social categorization

78 Rethinking Ethnicity

than of group identification (although still a matter of both). Third, while ethnic identity is, of course, part of a structured or coherent body of knowledge about the human world – as an aspect of culture, how could it be otherwise? – ‘racial’ categorization often appears to be both more explicit and more elaborated in its justification. Finally, and to repeat this chapter’s point of departure, regardless of whether the comparative perspective is historical or cross-cultural, ethnicity seems to be a ubiquitous social phenomenon; those situations which we describe as ‘race relations’ are not.

The last two points – the role of explicit justificatory knowledge, and the greater or lesser ubiquity of each – raise slippery questions about the respective ontological statuses of ethnicity and ‘race’, questions which lurk unacknowledged within many discussions of the problem. These are questions that are begged by the physical– cultural distinction. Furthermore, operating with a contrast between ‘the physical’ and ‘the cultural’ risks sneaking into our discussions through the back door other, even more troublesome, conceptual oppositions, such as ‘objective’ versus ‘subjective’, or ‘material’ versus ‘ideal’. This may result in confusion, at best, and conflict about the nature of the topic under discussion, at worst.

In order to begin the work of clarification, I want to suggest that ethnic identification, although it is every bit (and only) a social and cultural construction, should be conceptualized as a basic or first-order dimension of human experience. There are three, rather different, aspects to this claim. The most general concerns the ubiquity of ethnic attachments. If we define ethnicity very broadly – as collective identification that is socially constructed with reference to putative ‘cultural’ similarity and difference – then it seems uncontroversial to suggest that this has probably been around for as long as cultured humans have lived in social groups. This is, however, a point of view that requires considerable care in its articulation:

Classifying the most diverse historical forms of social identity as ‘ethnic’ creates the scientifically questionable but politically useful impression that all ethnicities are basically the same and that ethnic identity is a natural trait of persons and social groups … This is not an argument which bears up to historical scrutiny. Rather, it is a nominalist operation intended to provide scholarly legitimation for ethnonationalist ideologies. (Lentz 1995: 305)

This warning is important, timely, and well taken. There is a danger, in adopting too general a definition of ethnicity, of neutering the concept, while at the same time according it dubious extra-social potency. But to say that ethnicity seems to be one of the ‘givens’ of culture and human social life should not be misconstrued as saying that it is ‘natural’, other than in the sense that culture is a ‘natural’ propensity of humans, which, in turn, is to say little beyond problematizing the value of the culture-nature distinction (Jenkins 2002a: 111–38). To illustrate the point by analogy, the basic realities of sociality (Carrithers 1992) and the close caretaking of infants and children are ubiquitous in human experience. This doesn’t mean – although it can easily be construed thus – that there are universal, and normative, patterns or forms of either.

In the same way, to argue that collective identifications and attachments based on perceived cultural differentiation appear to be historically ubiquitous is not to justify ethnic or nationalist excesses, along the lines of ‘they – or we – can do no

Ideologies of Identification 79

other’. Such an argument is vulnerable to politicized misrepresentation: a good example of how this can happen is the way in which Geertz’s original argument about ‘primordial attachments’ has been academically caricatured, as discussed in Chapter 4. But while the danger of rewriting history by definitional fiat – and lending support to arguments that are not only repugnant but certainly untrue – is omnipresent, and demands that we exercise vigilance, it is not inevitable. Ethnic attachments are not necessarily malign; what matters is that we should try to understand the circumstances in which they become so.

And there is another, arguably worse, danger: of overlooking the consistency over time of the principles of collective identification and affiliation with which we are concerned and thus, in the process, misunderstanding them. Ernest Gellner – whose hostility towards nationalism can scarcely be in doubt, and who knew what its costs were – recognized this in his discussion of the loyalty that people feel for the groups to which they belong and the fact that they identify with them:

If one calls this factor, generically, ‘patriotism’, then … some measure of such patriotism is indeed a perennial part of human life … nationalism is a very distinct species of patriotism. (1983: 138)

If we do not acknowledge a very general, ideal-typical model of something that we can call ‘ethnicity’ – which, witness Barth’s ‘the social organization of culture difference’, for example, will necessarily be imprecise – how are we in principle to distinguish ethnic attachments from kinship, neighbourhood or organizational attachments, for example, and how, correspondingly, can we compare different ‘ethnic’ situations?

The second aspect of the claim that ethnicity should be understood as a basic dimension of social life alludes to taken-for-grantedness and embeddedness. Ask the question: what is it that constructs ethnic identification? At least part of the answer has got to be culture. While Hughes and Barth are absolutely correct to insist that cultural traits don’t constitute ethnic difference, Handelman (1977) and, more recently, Cornell and Hartmann (Cornell 1996; Cornell and Hartmann 1998: 173–84) are equally correct to remind us that the ‘cultural stuff’ is not irrelevant either, a point of view that is also strongly implied by the ‘ethnicity as cognition’ arguments of Brubaker (2004: 64–87; Brubaker et al. 2004) and Levine (1999). And our culture – language, non-verbals, dress, food, the structure of space, etc. – as we encounter it and live it during socialization and subsequently, is for us simply something that is. When identity is problematized during interaction across the boundary, we have to make explicit – to ourselves as much as to Others – that which we have hitherto known without knowing about. This is what Bourdieu (1990: 53–97) calls habitus, the embodied and unreflexive everyday practical mastery of culture: unsystematic, the empire of habit, neither conscious nor unconscious. Nothing could be more basic and nothing more inextricably implicated in ethnicity (Bentley 1987).

In the third place, however, a great deal turns upon local histories, circumstances and situations. This, again, is to return to an earlier thread of my argument: ethnic attachments do not have the same salience and force everywhere, or for everyone. Experience, if nothing else, insists that this is the case. For many people(s), ethnicity is a background factor, part of the cultural furniture of everyday life, and

Ideologies of Identification 81

bone structure and so on … What exists is not ‘race’ but phenotypical variation: ‘race’ is a word used to describe or refer to such variation. (Miles 1982: 20)

The relationship between ‘race’ and phenotype cannot possibly be this neat. For example, visible phenotypical differences – such as hair colour, perhaps – may be invoked in ethnic identification without ideas about ‘race’ also being invoked. On the other hand, putative ‘racial’ differentiation may be categorically asserted on the basis of physical differences that are either invisible to the unassisted naked eye – cranial indices are an obvious example here – or utterly imaginary. Finally – and Miles seems, paradoxically, to recognize this himself when he says that one does not exist but the other does – phenotype and ‘race’ are different orders of thing: phenotype is the material product of the interaction of genetic endowment (genotype) and environment; ‘race’ is a cultural fiction.

Other authors can appear to be just as confused. Thus Wade, criticizing ‘the seemingly unexceptional approach’ of Banton and Rex, argues that their work,

poses it [i.e. phenotype] as an obvious objective biological fact when in fact it is a highly socially constructed one. For it is not just any phenotypical variation that has become racialised (in any of the changing definitions of race), but a specific set of variations that have become salient in long-term colonial encounters. (Wade 1993: 21)

Phenotype cannot, on the one hand, be ‘highly socially constructed’ and, on the other, the range of physiological variations – as implied by ‘not just any phenotypical variation’ – out of which specific sets are historically selected for attention. Wade has subsequently tried to clarify this issue, but he still leaves us with a tension between the socially constructed ‘phenotype that is taken to underlie race’ and phenotype as ‘all aspects of appearance – actually it includes even more than that’ (Wade 2002: 4). At best, one appears to be more socially constructed than the other. Perhaps the real message here is that the concept of the ‘phenotype’, given its biological origins and meaning, does not have anything useful to offer the social science of ‘race’.

While I might not want to go all the way with Banton in arguing that ‘there is no distinctive class of social relations to be identified as racial relations’ (1986: 49), his viewpoint resonates with that which I am suggesting here. To express that point of view in another way, ethnicity and ‘race’ are different kinds of concept, they do not actually constitute a true pair. To oppose the one to the other does not, therefore, seem to make much sense. The most that can be said is that, at certain times and in certain places, culturally specific notions of ‘race’ – or, more correctly, of ‘racial’ differentiation – feature, sometimes very powerfully, in the repertoire of ethnic boundary-maintaining devices; for a recent study of the racialization of contemporary Welsh identity, for example, see the work of Scourfield et al. (2006; Scourfield and Davies 2005). To anticipate my argument below, and in general agreement with Robert Miles (1989), it is racialization and racism – or, to anticipate myself even further, racisms – to which we should be attending, rather than ‘race’.

Does it matter that we should agree on the nature of the difference between ‘race’ and ethnicity? Perhaps not, but it is certainly more than a ‘quibble’, and it is important to attempt to be clear about the matter. The issue has probably been responsible for even more confusion, and generated greater imbalances between heat and